The Question - "Which engine should I use?"
by Dark Tengu · 03/24/2009 (1:53 pm) · 30 comments
I am confounded by the amount of engines that are now available to the indie game developer. We now have so many options (much more than when I purchased TGE). Some of our options are more legitimate than others. Some engines seem to be schemes just to take your money (Lawmaker...ouch). As I see it, the most viable options are (not in any particular order):
Torque 3D
Unity 3D
C4
Leadwerks Engine
Various OpenSource Engines
Torque 3D - I think this probably has the most overall promise of the engines. Torque 3D has promised a lot and set the bar high. For my next project Torque 3D is currently my favorite choice. However, looking at the Torque 3D page, T3D appears to be nothing more than some graphical enhancements, web publishing, and Collada. No information about what makes the new generation of editors any better than what they were (a new GUI doesn't count as a significant enhancement - IMHO). The Collada tests I ran didn't seem to be too successful. The web publishing is definitely a cool enhancement with A LOT of benefits to us indies. Torque already has many strengths (networking ;) ). I am cautiously approaching Torque 3D as the main choice for my project, I just need to have some real answers on the work flow besides shiny graphics (no matter how cool).
Unity 3D - Hands down the most artist friendly tool. No source code is a turn off and a turn on. I always hated the excuse on the GG forums of "You have the source, you can do it." I would much rather prefer to have all the systems exposed in script. Adding features though, completely depends on the Unity devs. The biggest turn off with Unity3D is the price, $1500. Right now, I could get Torque3D for $505, much more appealing. I have never used Unity and probably never will. It didn't perform all that well on my Windows PC.
C4 - Seems to be one of the most promising engines around. Great CLEAN code. A highly motivated dev with a very aggressive release cycle. I would probably go with C4, but for some reason, the rendering engine gives me a headache and makes me dizzy (lol). Great price point at $350.
Leadwerks - Absolutely the best looking of the above engines. Sandbox editor is cool, but has some bugs still needing to be resolved. The price point is great at $150. Major problem is no networking at this point. Programming for the engine is extremely simple. No source with the base license, source license available. Very fast release cycle. If it had networking, this would be the no brainer choice for my project.
OpenSource - Lots of options available. I particularly like Sauerbraten. The absence of a lot of static meshes in the sample levels makes me very skeptical of the the performance though. Ogre looks amazing but it would take a lot of work to implement all of the necessary systems (I'm mainly an artist). Lots of other options available.
Well let me know of your opinions. It seems to be a great day to be an Indie developer with options all over the place.
Torque 3D
Unity 3D
C4
Leadwerks Engine
Various OpenSource Engines
Torque 3D - I think this probably has the most overall promise of the engines. Torque 3D has promised a lot and set the bar high. For my next project Torque 3D is currently my favorite choice. However, looking at the Torque 3D page, T3D appears to be nothing more than some graphical enhancements, web publishing, and Collada. No information about what makes the new generation of editors any better than what they were (a new GUI doesn't count as a significant enhancement - IMHO). The Collada tests I ran didn't seem to be too successful. The web publishing is definitely a cool enhancement with A LOT of benefits to us indies. Torque already has many strengths (networking ;) ). I am cautiously approaching Torque 3D as the main choice for my project, I just need to have some real answers on the work flow besides shiny graphics (no matter how cool).
Unity 3D - Hands down the most artist friendly tool. No source code is a turn off and a turn on. I always hated the excuse on the GG forums of "You have the source, you can do it." I would much rather prefer to have all the systems exposed in script. Adding features though, completely depends on the Unity devs. The biggest turn off with Unity3D is the price, $1500. Right now, I could get Torque3D for $505, much more appealing. I have never used Unity and probably never will. It didn't perform all that well on my Windows PC.
C4 - Seems to be one of the most promising engines around. Great CLEAN code. A highly motivated dev with a very aggressive release cycle. I would probably go with C4, but for some reason, the rendering engine gives me a headache and makes me dizzy (lol). Great price point at $350.
Leadwerks - Absolutely the best looking of the above engines. Sandbox editor is cool, but has some bugs still needing to be resolved. The price point is great at $150. Major problem is no networking at this point. Programming for the engine is extremely simple. No source with the base license, source license available. Very fast release cycle. If it had networking, this would be the no brainer choice for my project.
OpenSource - Lots of options available. I particularly like Sauerbraten. The absence of a lot of static meshes in the sample levels makes me very skeptical of the the performance though. Ogre looks amazing but it would take a lot of work to implement all of the necessary systems (I'm mainly an artist). Lots of other options available.
Well let me know of your opinions. It seems to be a great day to be an Indie developer with options all over the place.
About the author
#2
03/24/2009 (4:11 pm)
Turn off motion blur in C4. :P The implementation currently in there needs some tuning.
#3
03/24/2009 (4:22 pm)
Im moving to unity.. it fits my need better as a hobyist
#4
03/24/2009 (5:20 pm)
@Ron - I'm assuming your not using the professional license then. The Indie license is as worthless as Torque3D's Indie license. 
#5
03/24/2009 (5:36 pm)
What my plans are is to start with TGB, get a couple of games out and continue working from there. I honestly think that GG should leverage it as a beginners series for getting people into creating and designing games.
#6
My next project will almost certainly be based on T3D, though I will be keeping an eye on C4. So far I'm not overly impressed with their demo, though I've heard a lot of good things about the code base. Simple to get started doesn't equal simple to finish though, and I don't see anything released with it other than some subway simulator, so I'll wait and see what people do with it. It would be very hard to win me over from Torque, but I won't rule it out entirely.
The other commercial engines are non-starters because source code access is 100% necessary for me, and the 'contact us' pricetag is generally sales speak for 'you can't afford it if you don't have a NASDAQ symbol.'
03/24/2009 (5:45 pm)
Personally for my current project I'm sticking with TGEA. The new features in T3D don't justify the amount of work it will be to port everything over, and TGEA is a fantastic engine once you work around the few sore spots. Some of the code and art packs I'll be releasing will support T3D, but the primary game project is going to be TGEA (though a heavily modified TGEA).My next project will almost certainly be based on T3D, though I will be keeping an eye on C4. So far I'm not overly impressed with their demo, though I've heard a lot of good things about the code base. Simple to get started doesn't equal simple to finish though, and I don't see anything released with it other than some subway simulator, so I'll wait and see what people do with it. It would be very hard to win me over from Torque, but I won't rule it out entirely.
The other commercial engines are non-starters because source code access is 100% necessary for me, and the 'contact us' pricetag is generally sales speak for 'you can't afford it if you don't have a NASDAQ symbol.'
#7
It has it's own set of issues just like other engines, but Panda3D may be worth a look.
BSD-style, open-source license (as opposed to GPL/LGPL). Free. Pretty much everything that is inside the C++ engine is exposed to scripting (Python) via a very nifty bit of software tech. I'd say it's more programmer-friendly than artist-friendly, if Unity is at or near the top of the artist-friendly scale.
03/24/2009 (6:56 pm)
Quote:I would much rather prefer to have all the systems exposed in script.
It has it's own set of issues just like other engines, but Panda3D may be worth a look.
BSD-style, open-source license (as opposed to GPL/LGPL). Free. Pretty much everything that is inside the C++ engine is exposed to scripting (Python) via a very nifty bit of software tech. I'd say it's more programmer-friendly than artist-friendly, if Unity is at or near the top of the artist-friendly scale.
#8
i have 2 folders with engines one is a list of ok engines and the other is a list of over 200 bad engines. one of them is not a ok engine but is in the folder anyway as i like it the most out of any engine and thats HeroEngine. HeroEngine cost $300k for the first license all the way up to $950k for there last engine price so its out of a my reach.
if you want me to list engine sites here i can do that but it will take me some time to seet all them up. :P
03/24/2009 (8:17 pm)
i have looked at over 200 engines i took every single one out because they just was not that good. i have picked TGEA/T3D for many reason one is the community 2 is that GG is finally sticking with 1 3D engine for life which is T3D. theres many assets already done for torque they will be updating the engine for as long as GG still exists.i have 2 folders with engines one is a list of ok engines and the other is a list of over 200 bad engines. one of them is not a ok engine but is in the folder anyway as i like it the most out of any engine and thats HeroEngine. HeroEngine cost $300k for the first license all the way up to $950k for there last engine price so its out of a my reach.
if you want me to list engine sites here i can do that but it will take me some time to seet all them up. :P
#9
I know that I am going to loose my pre-order discount, but that is just something I will have to factor in at a later stage. My reasoning for purchasing T3D has nothing to do with it being better or not ... it is becoming the flagship 3D engine of GG and therefore I will purchase it.
So for the forseeable future I will still be working on TGEA with the forward thought of moving to T3D in a few months time once I have finished my current projects on TGEA. :)
03/25/2009 (2:01 am)
Although I will be purchasing T3D, for now I am also sticking with TGEA. Soul Wars - Lite will be released on TGEA 1.7.1 (AFX 1.1.2) with the final game going out on TGEA 1.8.1. I have invested too much time and money into TGEA and Soul Wars to warrant the upgrade cost to T3D right now.I know that I am going to loose my pre-order discount, but that is just something I will have to factor in at a later stage. My reasoning for purchasing T3D has nothing to do with it being better or not ... it is becoming the flagship 3D engine of GG and therefore I will purchase it.
So for the forseeable future I will still be working on TGEA with the forward thought of moving to T3D in a few months time once I have finished my current projects on TGEA. :)
#10
The market is interesting and if you like technology then it is very exciting.
I think the final decision is a matter of taste and depends on the project you are making.
For small 2d projects I think TGB is the best choice followed by Unity. For small casual games with optional 3d rendering Unity will be the best workflow, if you want to finish them fast. For bigger multiplayer games you could modifiy any engine or use Torque (maybe T3D).
But if you want to get the latest technology then C4 is worth a look, you have all the editors built in. You can edit models, world, materials, visual node based shader, terrain, visual scripts and more. You can create voxel terrain (like a 3d sandbox, no heightmap, you can sculpt in every direction and create steep walls, caves and more).
You have fantastic scene management: zones and portals for indoor, occlusion portals for blocking geometry, octree and grouping of meshes for outdoor and soon you get the world's first voxel terrain LOD system for really big outdoors.
This means: you can render millions or more polygons indoor and outdoor, you can have the highest details in all indie engines, the best visuals if you optimize the level and if you use the portals right.
So for an ambitious project I would (and I did) choose C4.
As I said, it depends on your project.
At Dexsoft we are creating models and textures for sale and for customers and we test them in several engines. I like the Torque renderer better than Gamestudio as an example. I like Unity because of the easy setup of shaders and shadows, but it always looks best in C4, I simply have more options to setup a material.
03/25/2009 (4:37 am)
I perfectly understand the situation of Marcus, Dale and others.The market is interesting and if you like technology then it is very exciting.
I think the final decision is a matter of taste and depends on the project you are making.
For small 2d projects I think TGB is the best choice followed by Unity. For small casual games with optional 3d rendering Unity will be the best workflow, if you want to finish them fast. For bigger multiplayer games you could modifiy any engine or use Torque (maybe T3D).
But if you want to get the latest technology then C4 is worth a look, you have all the editors built in. You can edit models, world, materials, visual node based shader, terrain, visual scripts and more. You can create voxel terrain (like a 3d sandbox, no heightmap, you can sculpt in every direction and create steep walls, caves and more).
You have fantastic scene management: zones and portals for indoor, occlusion portals for blocking geometry, octree and grouping of meshes for outdoor and soon you get the world's first voxel terrain LOD system for really big outdoors.
This means: you can render millions or more polygons indoor and outdoor, you can have the highest details in all indie engines, the best visuals if you optimize the level and if you use the portals right.
So for an ambitious project I would (and I did) choose C4.
As I said, it depends on your project.
At Dexsoft we are creating models and textures for sale and for customers and we test them in several engines. I like the Torque renderer better than Gamestudio as an example. I like Unity because of the easy setup of shaders and shadows, but it always looks best in C4, I simply have more options to setup a material.
#11
03/25/2009 (11:23 am)
Well, I decided to go with C4. I changed the FOV in the config file and it made all the difference. The cost is great, free updates for life. Torque3D ended up being what I expected for TGEA. I don't want to pay an additional $505 for what I should have gotten a long time ago. The lack of the full rewrite of Torque is what ultimately turned me off. I'm sure GG will want another $200 for the Torque3D 2.0 a year from now. Adios GG! ;)
#12
Otherwise, maybe you should wait for Torque X 3D to be released, you'll have to learn C# though.
03/25/2009 (5:39 pm)
Panda3D is indeed nice, from an experienced TorqueScripter perspective I found it quite easy to understand.Otherwise, maybe you should wait for Torque X 3D to be released, you'll have to learn C# though.
#13
03/25/2009 (8:12 pm)
ya you get C4 updates for life because it don't move that fast theres only 1 person working on it at lest the last time i heard about it there was only 1 working on that entire engine. the last time i heard about it was like 3 months ago i don't think anything changed. :P
#14
http://www.terathon.com/c4engine/notes.php
The engine is very nicely coded, which is something Torque lacks. Eric does a good job of actually following along with his roadmap for future updates as well.
http://www.terathon.com/wiki/index.php?title=Official_Roadmap
C4 requires more coding. If I were the OP, I'd look at Unity to prototype, then C4/T3D to actually make the game.
03/26/2009 (12:53 am)
Then you would be mistaken Mr. Baker. There have been multiple updates in the last 2 months.http://www.terathon.com/c4engine/notes.php
The engine is very nicely coded, which is something Torque lacks. Eric does a good job of actually following along with his roadmap for future updates as well.
http://www.terathon.com/wiki/index.php?title=Official_Roadmap
C4 requires more coding. If I were the OP, I'd look at Unity to prototype, then C4/T3D to actually make the game.
#15
You're right too, Eric does to a really good job of pushing updates often. For some people worried about stability, maybe too often. But, I think the point Brandon was making was that depending on one person to advance and support a technology is a bit more risky than an entire team. There's no question we'll be able to move faster than C4, but that doesn't mean that C4 hasn't come a long way or that it isn't solid, serious technology. I would like to see some good, commercial projects come from that technology and out of that community to really prove it out. It seems like Eric is targeting an interesting feature set.
03/26/2009 (1:14 am)
@Jason:Quote:If I were the OP, I'd look at Unity to prototype, then C4/T3D to actually make the game.While I agree that this would be the smart use of each engine if all 3 were already owned, I'd recommend choosing a product that you were confident of being able to get you to the finish line.
You're right too, Eric does to a really good job of pushing updates often. For some people worried about stability, maybe too often. But, I think the point Brandon was making was that depending on one person to advance and support a technology is a bit more risky than an entire team. There's no question we'll be able to move faster than C4, but that doesn't mean that C4 hasn't come a long way or that it isn't solid, serious technology. I would like to see some good, commercial projects come from that technology and out of that community to really prove it out. It seems like Eric is targeting an interesting feature set.
#16
the kits that will be released it will be real good. i looked into the forest kit some more on the guys site looks like a well worth it kit to me. to bad theirs not much info on his site ether as for the other kits cant really see any info on those i like to see if the adventure kit will be better then ubiq visuals kit for TGEA as theres is way over prices still and they not even going to change it they said even after 20 people complained about the pricing. they not even going to change how there license is set up ether where if your ready to release you have to pay 5k more just to release your game with the kit ya like who's going to do that the kit is not even that good as all that stuff has already been made just about.
i still going to with with T3D its well worth the price tag anyway for what you get.
went little off topic sorry. :P
Edit: i agree what Brett Seyler said as that is what i think though i did not look at the site still after i posted that other post. i am amazed he can do all that at that kind of speed but i think he will slow down at some point as he will start getting into more complex stuff even for him self hes going the easy stuff so far. i thinks hes also looking for free resources wiles hes doing all this so he don't have to do all that work.
03/26/2009 (1:16 am)
ya i seen those i bet the performance on C4 sucks as well that should be one big thing people should look at if your thinking of making a MMO which seems most people do. T3D is far ahead of C4 though when it comes out might not be in every area though but its still alot better and they have nto even shown everything for T3D yet. i still hoping there next blog is goign to be physics its goign to blow peoples minds though they said the physics is not the main thing they worked on but it still still be real super good.the kits that will be released it will be real good. i looked into the forest kit some more on the guys site looks like a well worth it kit to me. to bad theirs not much info on his site ether as for the other kits cant really see any info on those i like to see if the adventure kit will be better then ubiq visuals kit for TGEA as theres is way over prices still and they not even going to change it they said even after 20 people complained about the pricing. they not even going to change how there license is set up ether where if your ready to release you have to pay 5k more just to release your game with the kit ya like who's going to do that the kit is not even that good as all that stuff has already been made just about.
i still going to with with T3D its well worth the price tag anyway for what you get.
went little off topic sorry. :P
Edit: i agree what Brett Seyler said as that is what i think though i did not look at the site still after i posted that other post. i am amazed he can do all that at that kind of speed but i think he will slow down at some point as he will start getting into more complex stuff even for him self hes going the easy stuff so far. i thinks hes also looking for free resources wiles hes doing all this so he don't have to do all that work.
#17
Of course that does bring into question the future of it. And I believe that TGE promised lifetime updates too originally, and obviously that has changed, so I would take that promise with a grain of salt; licenses are always subject to change. There is probably going to come a point when he's not bringing in enough new users to justify continuing free updates on it, and he'll have to either charge again or close up shop.
It does indeed seem to be an ambitious feature set though. If he can pull it all off and deliver on the promises then that would be outstanding.
03/26/2009 (1:32 am)
Yeah I have to say I'm fairly impressed that C4 is a one-man show, I did not know that before.Of course that does bring into question the future of it. And I believe that TGE promised lifetime updates too originally, and obviously that has changed, so I would take that promise with a grain of salt; licenses are always subject to change. There is probably going to come a point when he's not bringing in enough new users to justify continuing free updates on it, and he'll have to either charge again or close up shop.
It does indeed seem to be an ambitious feature set though. If he can pull it all off and deliver on the promises then that would be outstanding.
#18
1) The license will not change for existing customers. This has been asked and answered a lot in the C4 forums. Customers get free upgrades forever.
2) "he will slow down at some point as he will start getting into more complex stuff"
If you write this then you have no idea about what Eric does. He wrote a voxel based terrain in a very short time, now he is finishing the world's first LOD system for such a voxel terrain. He wrote a node based shader editor in a few weeks. The tools are all in. A new job manager class has been introduced just as a little minor update to support multi-processor environments, ...
3) "There's no question we'll be able to move faster than C4.."
Hm, this sounds funny to me because I watch both technologies for a long time. C4 growth has been much much faster all the time. New tools came just at the same time as the underlaying features came, like a little add-on, like a free gift to use the feature. GG tools came from customers, late, long development cycles, bugs, problems to transfer the contents into the engines, restrictions, not up to date.
The same counts for shadows, shaders whatever. I still cannot edit materials in Torque or create shadows for every mesh in an easy way.
4) C4 performance: There is a wrong picture in some heads about C4 performance. You have portals and zones for indoors plus occlusion geometry to prevent everything behind from rendering. Outdoors use grouping and octree and soon the new lod system to have endless terrain. Background streaming has been announced to top this.
Even dynamic shadows use techniques to be optimized, how they will be seen through portals or how they behave if they dont move.
Actually you can render millions of polygons in a level, even every room can have something around that. You can bring it down with tons of dynamic lights though, but this is a hardware issue with every technology.
With all this in mind I ask myself whether a small team can be faster than a big one? It looks like this for me. Maybe the communication in bigger teams adds alot of time. And maybe it is hard do coordinate, to keep the same quality level among developers, to find bugs and to integrate new modules.
03/26/2009 (3:46 am)
There are a lot of false informations in this thread regarding C4. 1) The license will not change for existing customers. This has been asked and answered a lot in the C4 forums. Customers get free upgrades forever.
2) "he will slow down at some point as he will start getting into more complex stuff"
If you write this then you have no idea about what Eric does. He wrote a voxel based terrain in a very short time, now he is finishing the world's first LOD system for such a voxel terrain. He wrote a node based shader editor in a few weeks. The tools are all in. A new job manager class has been introduced just as a little minor update to support multi-processor environments, ...
3) "There's no question we'll be able to move faster than C4.."
Hm, this sounds funny to me because I watch both technologies for a long time. C4 growth has been much much faster all the time. New tools came just at the same time as the underlaying features came, like a little add-on, like a free gift to use the feature. GG tools came from customers, late, long development cycles, bugs, problems to transfer the contents into the engines, restrictions, not up to date.
The same counts for shadows, shaders whatever. I still cannot edit materials in Torque or create shadows for every mesh in an easy way.
4) C4 performance: There is a wrong picture in some heads about C4 performance. You have portals and zones for indoors plus occlusion geometry to prevent everything behind from rendering. Outdoors use grouping and octree and soon the new lod system to have endless terrain. Background streaming has been announced to top this.
Even dynamic shadows use techniques to be optimized, how they will be seen through portals or how they behave if they dont move.
Actually you can render millions of polygons in a level, even every room can have something around that. You can bring it down with tons of dynamic lights though, but this is a hardware issue with every technology.
With all this in mind I ask myself whether a small team can be faster than a big one? It looks like this for me. Maybe the communication in bigger teams adds alot of time. And maybe it is hard do coordinate, to keep the same quality level among developers, to find bugs and to integrate new modules.
#19
Saying it doesn't make it true. This is one of the most violated promises in the industry (not just game engines, but software development packages in general).
It's not made with treacherous intent, but with naivety of the hard reality of business. Nobody can support a commercial product with indefinite updates for life with any kind of aggressive update cycle unless it's a product that's going to generate a steady, continuous stream of new customers to support the old customers, which is not something you can count on with a game engine, especially a startup. Everybody has to eat and pay bills, and I doubt this Eric is any exception, otherwise C4 could just be free.
03/26/2009 (4:18 am)
Quote:
The license will not change for existing customers. This has been asked and answered a lot in the C4 forums. Customers get free upgrades forever.
Saying it doesn't make it true. This is one of the most violated promises in the industry (not just game engines, but software development packages in general).
It's not made with treacherous intent, but with naivety of the hard reality of business. Nobody can support a commercial product with indefinite updates for life with any kind of aggressive update cycle unless it's a product that's going to generate a steady, continuous stream of new customers to support the old customers, which is not something you can count on with a game engine, especially a startup. Everybody has to eat and pay bills, and I doubt this Eric is any exception, otherwise C4 could just be free.
#20
Besides that there is a commercial license for big companies and they have to pay per title anyway.
As far as I know, the cash flow was never a problem for Terathon.
03/26/2009 (4:34 am)
Yes, but this is no problem. The pricing will change for upcoming versions and the license for new customers might change as well. But old customers still get upgrades for free.Besides that there is a commercial license for big companies and they have to pay per title anyway.
As far as I know, the cash flow was never a problem for Terathon.
Torque Owner Dale Cannon